390 Analyst, June, 1973, Vol. 98, p p . 390-411 Pyridylazonaphthols (PANS) and Pyridylazophenols (PAPS) as Analytical Reagents Part 11." Spectrophotometric and Solvent-extraction Studies of Complex Formations BY D. BETTERIDGE AND D. JOHN? (Chemistry Department, University College of Swansea, Swansea, Glamorgan, SA2 8PP) The reactions between 2-(2-pyridylazo)-l-naphthol (o-oc-PAN), 1-(2- pyridylazo)-2-naphthol (o-P-PAN) and 2-( 2-pyridy1azo)phenol (0-PAP) with manganese(II), zinc(I1) and lanthanum(II1) and 4-(2-pyridylazo)phenol (#-PAP) with cobalt(II), nickel(I1) and copper(I1) have been studied. The spectrophotonietric procedure based on linear extrapolation, as used by Sommer, has been critically evaluated. The results from the spectrophoto- metric method have been used to predict the optimum conditions for solvent extraction.It is shown that this procedure is a valuable approach for systems in which hydroxy-complexes are common. SEVERAL studies have shown that o-P-PAN is very useful for the extraction of many cations.1-3 However, the extraction systems are often more complex than some of the early work suggests, and at present few detailed solvent-extraction studies have been made. Preliminary studies showed that 2-(2-pyridylazo)-l-naphthol (0-a-PAN) is often more advantageous than 1-(2-pyridylazo)-2-naphthol (o-P-PAN) and that 2-(%pyridylazo)phenol (0-PAP) and 4-(2- pyridy1azo)phenol (@-PAP) also form extractable complexes. The conventional method of determining equilibrium constants from solvent-extraction experiments is time consuming. It is possible, in principle, to determine the complex formation constants by other means, to determine the partition coefficient of the reagent and complex experimentally and to use the values obtained to predict the extraction graphs.One practical difficulty is that the con- stants that are obtained from solvent-extraction data relate to the aqueous phase saturated with the organic solvent and as, almost by definition, the complex is only sparingly soluble in water, the equilibrium constants must be determined in some other medium, e.g., 1 + 1 dioxan - water, so that they cannot be used directly for predicting extraction equilibria. Spectrophotometric methods can partially overcome this difficulty because solutions can be used that are so dilute that the complex can be maintained with such a small proportion of organic constituent that the values of the equilibrium constants are close enough to those obtained by solvent-extraction procedures to be interchanged.However, most spectrophoto- metric methods are based on Job's method, which can result in misleading or erroneous results, or both,4 and the study of a system over a wide range of conditions with these methods is extremely tedious. Recently, Sommer and co-workers5-11 have demonstrated that spectrophotometric procedures based upon linear extrapolation, when used with care, can be used most advan- tageously to study the complex equilibria of systems based on compounds analogous to the compounds discussed below. A practical advantage of their approach is that it is based upon the analysis of pH - absorbance curves so that no superfluous information is gathered and basic data are interpreted fully.In this paper, we assess this procedure by analysing several systems, for some of which results are already available for comparison, and show how it can be used for predicting extraction curves with an accuracy that is acceptable to the analyst. EXPERIMENTAL REAGENTS- Salts and solvents of analytical-reagent grade purity or better were used throughout. * For Part I of this series, see p. 377. t Present address: BP Chemicals (U.K.) Ltd., Llandarcy, Swansea. @ SAC and the authors. Parts I11 and IV will appear in the July issue.BETTERIDGE AND JOHN 391 1-(2-Pyridylaz0)-2-naphthol (0-p-PAN). 2- (2 - Py ridy 1 azo ) - 1 -naphthol ( o-a- P A N ) .2- (2-Pyridylazo)phenol (0-PA P) . 4-(2-Pyridylazo)phenol (p-PAP) . The above four reagent solutions were prepared and purified as described in Part I. Solutions in absolute or aqueous ethanol were prepared for spectrophotometric studies and in carbon tetrachloride or chloroform for solvent extraction. The concentrations are given in the procedures. The solutions were taken to be standard and the validity of this assumption was checked occasionally by spectrophotometric titration of a standardised metal-ion solution. Metal-ion solutions-Solutions of manganese(II), zinc(II), nickel(II), cobalt(I1) and lan- thanum( 11) were prepared and standardised titrimetrically with EDTA by standard procedures. Bufer solutions-Standard chloride, phthalate, phosphate, borate and hydroxide buffer solutions to cover the pH range from 0 to 12 were used.Sodium perchlorate-Solid sodium perchlorate was used in order to maintain an ionic strength of 0.10 & 0.01. Radioisotopes-Manganese44 and zinc-65 were obtained from the Radiochemical Centre, Amersham. They were diluted and mixed with carrier manganese(I1) and zinc(I1) so that solutions of known concentration and radioactivity were obtained. APPARATUS- pH meter-Radiometer, Model M4C. Spectro$hotometers-Unicam SP500, SP600 and SP800 and Cary, Model 16, instruments were used as appropriate. Absorbance values used for the calculation of equilibrium constants were obtained on fixed-wavelength instruments. Radioactivity-A 1-inch well sodium chloride (thallium-activated) crystal connected to a photomultiplier and a 1DL 1700 scaler was used to measure the radioactivity.Samples of constant volume were taken so as to ensure constant geometry and sufficient counts were taken so as to ensure a statistical error of not more than 1 per cent. on all except the very low count-rates, when an error of 10 per cent. was accepted. Computer-An IBM 1600 computer was used for the calculation of constants based on the spectrophotometric data and the statistical analyses of them. The standard program for linear regression in two variables 1620/6.0.27 from the 1620 General Program Library was used. PROCEDURES- Spectrophotometric determination of acid-dissociation constants-A 10-ml aliquot of a solution of o-a-PAN (1 x M) or $-PAP (1.25 x 10-4~) in 12.5 per cent.aqueous ethanol was placed in a 25-ml calibrated flask and sufficient solid sodium perchlorate to maintain an ionic strength of 0.10 & 0.01 was added. The solution was then made up to the mark with buffer solution. The solution was mixed well, the spectrum recorded and the absorbance and pH were measured. A pH - absorbance curve was plotted for each of the reagents at suitable wavelengths and the pK, values were calculated from the inflection point as determined from a graph of the differential AAIApH against pH. The pK, values were also calculated from the same pH - absorbance curve by the procedure described below. Spectrophotometric determination of formation constants-pH - absorbance curves were obtained with both the metal ion and the reagent in excess.The general procedure was to place 10 ml of ethanolic solution, 10 ml of metal-ion solution and 5 ml of buffer in a 25-ml calibrated flask that contained sufficient solid sodium perchlorate to maintain the ionic strength of the final solution at 0-10 -+ 0.01. These solutions were mixed thoroughly and the colour was allowed to develop to a maximum (a few minutes). The absorbance remained constant for the period of the measurement and showed little change after 24 hours. The spectra were recorded, the absorbances were determined at suitable wavelengths and the pH values were measured. The measurements were carried out a t room temperature. The method of continuous variations12 and the mole-ratio method13 were also applied to some systems.Solvent extraction and determination of acid-dissociation constants-The reagent solutions were made up in carbon tetrachloride (for o-a-PAN, o-p-PAN and o-PAP) or chloroform (for $-PAP) to concentrations of 1.0 x 1.0 x 2.0 x and 5.0 x 1 0 - 4 ~ ~ One-centimetre cells were used. M) in absolute ethanol or o-PAP (2.5 x392 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, Vol. 98 respectively (@-PAP is insufficiently soluble in carbon tetrachloride for the distribution ratio to be measured accurately). A 5-ml aliquot of reagent solution was placed in a vial and 60 ml of buffered aqueous phase of ionic strength 0.10 & 0.01 were added. The solutions were shaken overnight in a box shaker that was maintained thermostatically at 24 &- 1 "C. The phases were allowed to separate and the absorbance of the organic phase was measured with the Unicam SP600 instrument and that of the aqueous phase, after adjustment of the pH to zero with concentrated hydrochloric acid, with the Cary 16 instrument.Calibration graphs were used to convert the absorbance values into concentration values. The distribution ratio of the reagent, D, was calculated from D = (CT - c W ) v W / c W v O D = CoVw/CwV, for D > 1, or for D < 1, where CT is the total concentration of the reagent, C, and C, are the concen- trations of the reagent in the organic and aqueous phase, respectively, and V , and Vw are the volumes of the organic and aqueous phase, respectively. Log D was plotted against pH and the partition coefficient of the reagent, KDR, and acid-dissociation constants, Kal and K,,, were calculated from1* D = KDR { [HI/KaI + 1 + KaJ [HI 1 Solvent extraction of metal ions-Reagent solutions were 5.0 x M in carbon tetra- chloride.M were diluted to a suitable concentration when the extraction was followed spectrophotometrically. When the extraction was followed radiochemically, a solution that was as dilute as the specific activity of the isotope would allow was used and a 0.2 M solution of sodium perchlorate was prepared. Equal volumes of organic and aqueous phase were used, the latter consisting of 2 nil of metal-ion solution, 2 ml of buffer solution and 4 ml of sodium perchlorate solution. The solutions were placed in a vial and shaken overnight, the layers were separated and the radioactivity of each phase was measured.The total activity in the two phases was computed and if it was less than 80 per cent. of the total activity in the vial, the results were discarded. Stock metal-ion solutions of 5 x RESULTS AND DISCUSSION The following symbols are used- HR, H,R+, R- CCXI a, P f the neutral, protonated and ionised forms of the reagent, respectively the total analytical concentration of species x the molar absorptivity of species x the total absorbance IMlo/lMlw, the distribution ratio [HR],/ [HR], or [MRn]o/ [MRnIw, the partition coefficient of the reagent [HR] [H+]/ [H2R+], the first acid-dissociation constant [R-] [H+]/[HR], the second acid-dissociation constant [MR,]/[M] [R-I2, the over-all formation constant of MR2 and extractable complex, respectively [MRI / CMI [RI WR2I [H+l2/[M1 WRI2 WR2OHl / WR2I PHI [H,,R]~%~H,-,R] the fractional extent of completion of the reaction as written.x=o [MI /CMT0J Charges are omitted unless possible ambiguity would arise from their absence, REAGENT EQUILIBRIA IN THE ABSENCE OF METAL IONS SPECTROPHOTOMETRIC STUDIES- Principles of graphical analysis of absorbnnce versus pH curves-This method was re- examined by Sommer5 in 1964 and has since been applied to the study of a number of reagents.6~7 The principles involved are summarised below.June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 393 The method involves a detailed algebraic study of the equilibria of the reagent, and in the present study these are taken to be: Kaz Ka, HR + R- + H+ (Equilibrium 1) H,R+ + HR + H+ (Equilibrium 2) On the initial assumption that these equilibria represent the true reagent equilibria, a series of linear algebraic transformations can be derived. Experimental values of absorbance and pH are introduced into these transformations, which then provide values of the acid- dissociation constants, Kal and Ka,, and also molar absorptivity values of individual reagent species.An observed lack of linea.rity in these transformations indicates that the assumed equilibria are incorrect. In Equilibrium 1, if it is assumed that only the neutral (HR) and anionic (R-) forms of the reagent are present, it follows that the total reagent C R = [HRI + LR1 = [RI {e+ l} The total absorbance, A , is given by A = EHR [HRI + ER [R-I concentration, CR, is given by ..' (1) .. .. Substitute for [R] from equation (1) and re-arrange to give .. Multiply throughout by l/eR and re-arrange to give - (A - €HRCR) CR A =-+ €R Ka,ERA A graph of CR/A zleysus [H]A is usually a straight line of Values of the acid-dissociation constant, Ka,, can be obtained sponding values of absorbance and pH. Alternatively, it can be argued that r and A = [HR] [EHR + 6 3 1 [HI .. .. .. .. (1) .. .. intercept l/en, as A > EHRCR. by calculation by using corre- .. .. .. .. .. .. .. .. which give This equation, on multiplication by l/gHR and subsequent re-arrangement, gives .. .. CR - 1 Ka, ( A - CRER). 1 A -EG + EHR A[H] " Equation (IA) should also yield a straight line when CR/A is plotted against l/A[H]. The intercept of this line represents the reciprocal of the molar absorptivity, EHR.Values of Ka, can be calculated from corresponding values of absorbance and pH. Alternatively, a value of eR, which can be found from other experiments, can be substi- tuted and &/A plotted against ( A - CR~R)/A[H]. A similar procedure can be adopted to solve graphically all of the subsequent transformations, where the linearity is not obvious.394 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, vol. 98 These equations are therefore derived simply from consideration of mass balance and equilibrium constants, with the one important algebraic manipulation being effected by the multiplet ~ / E H R or ~/ER. In equilibrium 2, if it is assumed that [R-] is negligible compared with [H,R] and [HR], then and By using shown that CR = LHR1 + LHzR1 A = EHR[HR] + EH,R[H&] these equations and proceeding in the same manner as before, it can be and For equation (11), the graph of CE/A veysus l/A[H] should be a straight line.Similarly, the graph of CR / A versm [H]/A for equation (IIA) should be linear. Values of the first acid- dissociation constant, &, and of the molar absorptivities, EHR and c H 2 ~ , can be calculated from these equations. Absorbance curves and results of calculations-The absorbance veysus pH curves are shown in Fig. 1. The wavelength of maximum absorption for the various species and isosbestic points are given in Table I. The results obtained for each transformation are summarised in Table 11. The transformations are represented as straight-line equations of the form y = mx + c, where y = &/A and x = 1/A [HI or [H]/A multiplied by a power factor to keep x and y mainly within the range 0.1 to 10.The values obtained are in good agreement with those obtained from the inflection point of the pH - absorbance curve. It is slightly advantageous that the molar absorptivity is calculated simultaneously and that a check on the nature of the equilibrium is provided. If two protons were to be released simultaneously, for example, the observed transformations would be curved or extremely ill defined. The least-squares analysis permits the precision of the result to be calculated. It was found that because the method is based on an extrapola- tion, it is very sensitive to the choice of points used for the calculation.Inevitably, a t one end of the part of the pH - absorbance curve that is being used for analysis, the basic assump- tion that one species is of negligible concentration compared with the others becomes less valid. A point from this part of the curve will therefore be “bad” and because it will be at the end of the transformation it can exert a disproportionate effect on the least-squares analysis. Several checks are possible: (i) the point may be so “bad” that it falls outside the 95 per cent. confidence limits and can be rejected; (ii) the point can be discarded and the analysis carried out again, and if the result is then the same the point can be accepted and if it is markedly different it can be rejected; and (iii) a different transformation can be used and the results compared.Rejection by the second of these procedures is not entirely satisfactory and care was taken to reject not more than one result and to confirm, by examination of the experi- mental pH - absorbance curves, that this result was a marginal value. Checks of (i) and (iii) were always carried out. A practical disadvantage of the method, therefore, is that a larger number of “good” experimental points must be obtained than is necessary for the simpler spectrophotometric procedures. SOLVENT-EXTRACTION STUDIES- The experimental curves of log D versus pH were very similar to that already published for o-a-PAN.14 All of the curves indicated that a protonated species of reagent was formed under acidic conditions and an anionic species under alkaline conditions.The pK values obtained are given in Table 111. The logarithm of the partition coefficients for o-a-PAN, o-p-PAN, o-PAP and +-PAP are 5.20, 5.00, 3.75 and 2.53, respectively. It was found that although the partition coefficients for 0-a-PAN and o-P-PAN were an order of magnitude greater than those reported earlier, the acid-dissociation constants were in reasonable agree- ment. This agreement is surprising, as the values of the partition coefficients are used in the calculation of the acid-dissociation constants. With such high partition coefficients, theJune, 19731 1 1 - 1 1 1 1 1 1 1 1 1 1 1 1 1 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 0.9 0.8 0.7 395 (C) Ka 1 - HzR' + HR + H" HR i R-+ H+ pKaq= 2.70 - - 1 -2 1.0 08 0.6 0-4 0.2 2.0 0.4 0 PH Fig.1. Absorbance versus pH graphs: (u) 5.0 x M $-PAP in 6 per cent. aqueous ethanol; M o-a-PAN in 40 per (b) 1.0 x lo-* M o-PAP in 5 per cent. aqueous ethanol; and (c) 4.0 x cent. aqueous ethanolTABLE I ABSORBANCE MAXIMA, MOLAR ABSORPTIVITIES AND ISOSBESTIC POINTS H,R+ HR R- Isosbestic point A r \ Reagent wnm E )czlnm E &bm E H,R 4 HR+ + H+ HR + R- + H+ P-PAP 402 2-38 x 104 358 1.89 x 104 442 2-10 x 1 0 4 373 382 O-PAP 352 1.62 x 1 0 4 328 1-56 x lo4 480 9-16 x 103 337 410 o-a-PAN 365 1.63 x 104 486 1.62 x 1 0 4 520 2.08 x lo4 476 - TABLE I1 MOLAR ABSORPTIVITIES AND ACID-DISSOCIATION CONSTANTS OF REAGENTS FROM LEAST-SQUARES FIT OF TRANSFORMATIONS (I), (IA), (11) AND (IIA) Correlation Standard error of Acid-dissociation Molar absorptivity constant ER PKa, Transformation Reagent y = w + c coefficient I p-PAP y = 0.101x + 4.756 0.992 0.0046 0.183 2.10 x 104 7-66 f 0.02 0-PAP y = 0.023~ + 1.003 0.990 0*0010 0.057 9-16 x 103 8.77 f 0-02 o-a-PAN J.J = 0 * 0 8 1 ~ + 4.808 0.988 0.0045 0.048 2.08 x 10' 10.17 & 0.02 IA $-PAP 0-PAP o-E-PAN I1 P-PAP O-PAP o-a-PAN IIA P-PAP 0-PAP o-a-PAN y = 0.264~ + 5.056 y = 0.018~ -+ 6.638 - y = 0.183~ + 4.203 y = 0.030~ $- 6.189 y = 0.055~ + 6.154 y = 0.044~ + 5.561 y = 0.041% + 6.621 y = 0.099~ + 6.182 0.979 0.975 - 0.980 0.992 0-993 0.978 0.989 0.990 0.0192 0.0014 - 0.0142 0-0015 0.0020 0.0033 0~0020 0.0049 0.198 0.091 I 0.057 0.054 0.018 0.150 0.035 0-023 EH R 1.98 x 104 1-51 x 104 2.38 x 1 0 4 - EHIR 1.62 x 10' 1.63 x lo4 €HR 1-80 x 104 1.62 x 1 0 4 1.51 x lo1 PKa, 7.78 f 0.02 8.92 f 0.04 - PKa, 3.00 f 0.03 2.72 f 0.06 2-73 f 0.03 2.94 f 0.07 2.64 f 0.05 2.60 f 0.05 P h'a, No calculations were carried out for o-a-PAN by using transformation (IIX) because of the proximity of the Amax.values of the reagent species HR and R-. * Z U c-c 0 E * Z U n b x $June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 397 presence of a small amount (e.g., 0.1 per cent.) of a coloured impurity that is less extractable than the reagent can profoundly affect the experimentally determined value of the partition coefficient, but would scarcely affect the log D versus pH curve when the value of the dis- tribution ratio is less than 100. This effect would result in the experimentally determined pKa, value being greater than and the pKa, value being less than the true values.As there is an internal consistency in both sets of studies, which should prevent miscalculation, we can offer no explanation for the discrepancy unless there is some slight difference between the samples of reagent-grade carbon tetrachloride used in the two studies. The values obtained in this study were used throughout this work. TABLE I11 ACID-DISSOCIATION CONSTANTS AT AN IONIC STRENGTH OF 0.10 Reagent Method* O- p-PAN SP. Pot. S.E. S.E. o-E-PAN Pot. S.E. S.E. SP. SP- SP. SP. o-PAP SP. SP* SP. #-PAP SP. SP. SP. S.E. S.E. Medium 20% aqueous dioxan 50% aqueous dioxan Chloroform - water 50% aqueous dioxan Water Water - carbon tetrachloride Water - carbon tetrachloride 50% aqueous methanol 40% aqueous ethanol 40% aqueous ethanol 50% aqueous methanol 5 yo aqueous ethanol 5y0 aqueous ethanol Water - carbon tetrachloride 50% aqueous methanol 5% aqueous ethanol 5% aqueous ethanol Water - carbon tetrachloride - PK, 1.9 <2 2-9 2.9 2-54 3.0 3.1 2.90 2.29 2.67 2.70 1.85 2.68 2.68 2.68 2-47 2-97 2.86 3-58 PKa, 12.2 12.3 11.5 11.2 10.74 9.1 9.5 9.63 10.00 10.17 10.23 9.42 8.84 8.79 8.68 8-20 7-72 7.76 7-55 Reference 15 16 17 18 14 14 14 This work 19 This work, graphical This work, conventional 20 This work, graphical This work, conventional This work 20 This work, graphical This work, conventional This work * Sp.= spectrophotometric; Pot. = potentiomctric; S.E. = solvcnt extraction. COMPLEX FORMATION EQUILIBRIA SPECTROPHOTONETRIC STUDIES- The graphical method that is used for the determination of acid-dissociation constants can be applied in the study of the formation 01 complexes.The method consists in first postulating the chelation reaction and the species formed over a particular pH range, then algebraically deriving transformations that would necessarily be followed if the postulates were correct. Typically, these transformations are straight lines, and false assumptions are rcadily detected by the presence of curvature or random scatter when experimental values are substituted into the transformations. All parts of the simple absorbance - PI-I curve can be subjected to such analysis, so that the various conditions of chelation can be deduced over a wide pH range. Normally, absorbance-pH curves are obtained in the presence of excess of reagent and excess of metal ions, so that the absorbance can be expressed generally as A = f(PH) cma,cx A = .WM)~H,C~ A =~(CR> p ~ , cM Sommer and co-workers have summarised a large number of basic equations for a variety of chelation reactions in the presence of various concentrations of reagent and metal ions, and have since applied the method successfully to different systems.These systems include the reactions of 4-(2-thiazolylazo)resorcinol (TAR) and 4-(2-pyridylazo)resorcinol (PAR) with thallium and uranyl ions.8-f0 The reactions of copper(II), lead(II), cadmiuni(II), zinc(I1) and bismuth(II1) with TAR and of the lanthanides with PAR have also been reported.6911 and398 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [.4nd,St, VOl. 98 We have applied the method to the reactions of zinc(II), manganese(II), cobalt(II), nickel(II), lanthanum(III), copper(I1) and titanium(1V) with o-a-PAN, o-b-PAN, o-PAP and p-PAP and compared the results when possible with those obtained by different methods.Most of the findings are given below but some will be dealt with in subsequent papers. GRAPHICAL ANALYSIS OF ABSORBANCE CURVES- The general approach can be illustrated by the derivation of a few basic equations. One common reaction is chelate formation when the reagent is predominantly in the neutral form. The reaction can be represented as *KYR1 This reaction can be carried out with either metal ions or the reagent in excess and the un- reacted reagent may or may not contribute to the absorbance at the wavelength of maximum absorbance of the chelate.In the presence of excess of metal ions--In this instance, [MI = Cy, the total metal-ion concentration, and if it is assumed that E H R is negligible a t the wavelength being used, then Furthermore, if it is assumed that the concentration of the intermediate chelate species, MR, is negligible, then Re-arrange equation (5) and substitute for [MR,] : M+2HR + MR2+2H .. .. .. * * (3) The effects of these variations are considered below. .. * - (4) C R = [HR] + 2[MR,] . . .. .. - (5) A = EMR, [MRJ .. .. [HR] = CREMR, - ZA EMR, Divide equation (4) by equation (6) and substitute for [MR,]/[HR] by introducing the reaction constant, so that or Multiply through by ~ / A E ~ I R , and re-arrange : When the reagent makes a contribution to the absorbance, the expression for the total absorbance requires an additional term : A = EMR,[M&] EHR[HR] 6 - ... . . . (4a) The steps outlined above then give . . (IIIA) In the presence of excess of reagent-Under these conditions, [HR] = C,, the total reagent -- CR - - 1 . ( A - CREMK) . [HI2 A - EMR, + "KMR CMEMR, A [HR] ' ' concentration, and if the reagent absorption is negligible, then ' * (7) .. .. .. .. A [MR,] =- EMR, and if [MR] is negligible compared with [MI and [MR,], then so that C M = [RI] + [MR,] . . .. .. - - (8) Divide equation (7) by equation (9), substitute for [MR,]/[M] and re-arrange, to giveJune, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 399 Multiply through by ~/AEMR, and re-arrange, so that A- EMR, *KMR,c~REMR, [HI2 .. .... .. (IV) 1 1 CId - - + ~ - . When the reagent absorbance is not negligible, an extra term can be introduced. In practice, we found that either the reagent absorbance was negligible or it represented such a large contribution that internal compensations on the spectrophotometer had to be made. Another important type of reaction is the hydrolysis reaction KOH MR2 + OH- + MR,OH- .. .. .. . . (10) If it is assumed that the hydroxy-complex makes no contribution to the absorption, then in the presence of both excess of reagent and excess of metal ions the total absorbance, A , is given by A = EMR~ [MR2I For excess of metal ions, the total reagent concentration, CR, is given by For excess of reagent, the total metal-ion concentration, CM, is given by In the presence of excess of reagent: C R = 2[MR,] + 2[MR,OH] C M = [MI + PR2I + [MR,OHI .... - - (V) ions is academic because. or The derivation of equations in the presence of excess of metal in most systems, the metal hydroxide would be formed preferentially and be precipitated: The transformations for these and other systems are given in Table IV; some are in logarithmic form because this form is more convenient for use in calculations. All of the transformations can be expressed as a linear function, although some contain two variables, [HI and [HR], in one term. In these instances, as for example transformation (111), EMR, was calculated from another transformation, (IV), or by simpler conventional means, and this was used to calculate [HR] by means of equation (6). The calculation procedure is very similar to that used for the acid-dissociation constants and the same checks were applied.However, even greater care is necessary because although a pH - absorbance curve reflects the course of several reactions, which allows these reactions to be detected and studied, it is easier to introduce “bad” points into the calculations. A further check was therefore always carried out. The values of the calculated constants were used to derive a pH - absorbance curve and this curve was then compared with the experimental curve. Zinc(l1) chelates of o-a-PAN, o-P-PAN and o-PAP-pH - absorbance curves obtained for the chelates of each reagent in the presence of excess of metal ions and of reagent are shown in Fig. 2. The absorption maxima of the zinc(I1) chelate with o-a-PAN, o-P-PAN and o-PAP occurred at 548 and 590, 514 and 550, and 530nm, respectively, the chelates of the naphthol derivatives each having two maxima, and the pH - absorbance curves were obtained a t 590, 550 and 530 nm, respectively.These curves indicate that each chelate has a stoicheiometry of metal to ligand of 1 : 2. At the upper limits of pH, the absorbance values became inconsistent and no smooth curve could be drawn. Hence, over the pH range 6 to 8 in which the chelate is being formed, the basic reaction would appear to be Zn + 2HR + ZnR, + 2H Transformations (111) and (IV) were therefore used for the reactions with o-a-PAN and o-PAP and transformation (IIIA) was used for o-P-PAN, as the reagent contributed to the absorbance. The results are given in Table V, which shows that there is good agreement between the molar absorptivities and stability constants determined in the presence of both excess of reagent and excess of metal ions.The agreement between these values, together with the linearity of transformation (IIIA) and hence of transformation (VIA), confirm that the chelation reactionTABLE IV BASIC COMPLEX FORMATION REACTIONS AND LINEA.R TRANSFORMATIONS Principal reaction Condition M + ZHR+MR,+ZH Excess of metal ions Excess of reagent Excess of metal ions Excess of reagent MR, + OH+MR,OH Excess of reagent M + HRsMR + H Excess of metal ions Excess of reagent Excess of metal ions Excess of reagent M + 2R+MR, M + R+MR No. 111 IIIA IV VI VIA VII V VIII IX X X I W 2 4 M w Y U 0 M n b W M,June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS.PART I1 401 suggested is correct and that the assumptions made in the derivation of the equations are valid. Each chelate investigated therefore has a stoicheiometry of metal to ligand of 1 : 2, and is involved in a chelation reaction of the type PH Fig. 2. Absorbance zwsw pH graphs for zinc(I1) chelates: (a) in the presence of excess of metal ions- c M / M C R / M X/nm 0-a-PAN . . . . 2.0 x 1 0 - 4 4.0 x 10-6 590 0-P-PAN . . . . 2.0 x 1 0 - 4 5.0 x 10-5 550 O-PAP * . . . 2.0 x 10-4 6.0 x 10-6 530 c b f / M C R / M X/nm 0-M-PAN . . . . 2.0 x 10-6 2.0 x 10-4 590 o-P-PAN . . . . 2.0 x 10-6 2.0 x 1 0 - 4 550 0-PAP .. . . 2.0 x 10-6 2-0 x 10-4 630 (b) in the presence of excess of reagent- The values of the molar absorptivities and stability constants determined in this study were found to be consistent with the results of other Mangnnese(l1) chelates of o-a-PA N , o-P-PAN and o-PAY-The absorbance maxima of the nianganese(I1) chelates of o-a-PAN, o-P-PAN and o-PAP occurred at 646 and 686, 616 1 .o (a 1 5 6 7 8 9 1 0 1 1 1 2 PH Fig. 3.Absorbance versus pH graphs for manganese(I1) chelates: C m / h l C R / M h/nm (a) in the presence of excess of metal ions- 0-LX-PAN .. . . 2.0 x 10-4 4.0 x 10-6 686 0-PAP .. .. . . 2.0 x 10-4 4.0 x 10-6 630 o-P-PAN . . . . 2.0 x 10-4 4.0 x 10-6 546 (b) in the presence of excess of reagent- c a r / M Cn/M hlnm 0-LX-PAN .. . . 2.0 x 10-6 2.0 x 10-4 686 0-PAP .. . . .. 2.0 x 10-5 2.0 x 10-4 530 o - p - ~ ~ ~ .. .. 2.0 x 10-6 2.0 x 10-4 546 1 12TABLE V RESULTS FOR THE CHELATES OF ZINC(II) Standard error of -mate y = m t : + c coefficient slope, m Correlation y = 0.087~ + 4.515 1.000 0*0003 0-015 y = 0-804X + 8.349 1.000 0.0052 0.096 y = 0.066~ + 2.212 1.000 0.0002 0.016 y = 0.037% + 2.222 1.000 0-0002 0.0 15 y = 0.076~ + 4.165 1.000 0.0003 0.025 y = 0.453~ + 4.208 0.999 0.0055 0*108 pKa, = 10.20 (o-a-PAN); 12-20 (O-F-PAN); 8-80 (0-PAP).Condition Reagent Excess of metal ions o-a-PAN O- p-PAN 0-PAP Excess of reagent o-a-PAN O- 8-PAN O-PAP Condition Reagent Excess of metal ions o-a-PAN 0- 8-PAN 0-PAP Excess of reagent o-a-PAN 0- 8-PAN 0-PAP TABLE VI RESULTS FOR CHELATES OF MANGANESE(II) Standard error of Correlation r y = 0.619~ + 4.945 1.000 0.0006 0.01 1 y = 0.045~ + 2.455 1.000 0-000 1 0.009 y = m % + c coefficient slope, m estimate y = 0.874~ + 4.823 1.000 0.0006 0-029 y = 0.504~ + 9.201 1.000 0.0012 0.059 y = 0.086~ + 2.443 1.000 0-0002 0-015 y = 0.265% + 4.808 0.999 0-0032 0-132 pKa, = 10.20 (0-a-PAN) ; 12.20 (0-8-PAN) ; 8-80 (o-PAP).EMIR, x 4-43 4.75 2-40 4.52 4-50 2.40 EYR, x 4-15 4.04 2-17 4.09 4.07 2-08 log KMR, 19.13 f 0.07 21.63 f. 0.08 15-52 f 0.09 19.12 f 0.07 21.45 f 0.08 15-48 f 0.09 log KMR, 13-54 f 0.06 15-77 f 0.06 10.52 f 0.02 13.27 f 0.05 15-69 f 0.08 10-45 f 0.05 ?- z U 0 X Z n ? pJune, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 403 and 546, and 530 nm, respectively, the chelates of the naphthol derivative each having two maxima, and the pH - absorbance curves (Fig. 3) were obtained at 586, 546 and 530 nm, respectively.Over the pH range 6 to 9 the curves are similar in appearance to those found for the zinc(I1) chelates and the same analysis was applied. The results are shown in Table VI. At higher pH values, the absorbance of the neutral chelate was found to decrease with increase in pH. The decreased absorbance was usually accompanied by the formation of a precipitate, which was presumed to be a hydroxy-form of the chelate. With o-a-PAN and o-PAP, it was not found possible to correlate absorbance with pH, and it was observed that the absorbance tended to zero on standing the solution. The manganese(I1) chelate of o-P-PAN, however, gave reasonably smooth absorbance curves in the presence of both excess of reagent and excess of metal ions. The values tended to vary on standing the solution, but it proved possible to use them in transformation (V) (Fig.4). From these values, KMR,(OH) was found to be The linearity suggests that the hydrolysis reaction that was proposed is reasonable, but because of the variation noted above, we have some reservations about these conclusions. 0.2 - ( a ) 0 - -0.2 -0.4 - - 0.2 l,o.21 (b) ;l.l 0 -0.2 -0.4 -0.6 9.7 9.8 9-9 10.0 10.1 10.2 PH Fig. 4. Transformation (V) for the hydrolysis reaction of Mn(o-P-PAN),: (a) excess of reagent; and (b) excess of metal. Conditions as in Fig. 3 Lalzthanum(III) chelates of o-a-PAN, o-P-PAN and o-PAP-The chelation reactions of lanthanum( 111) with o-a-PAN, o-P-PAN and 0-PAP were investigated and absorbance curves in the presence of both excess of reagent and excess of metal ions were obtained.Absorbance measurements were made at 550 and 514 nm, these wavelengths correspond- ing to the absorbance maxima of the o-a-PAN and o-PAP chelates, respectively. The absorbance curves are given in Fig. 5 and show that the extent of chelation is greatly affected by hydrolysis of the lanthanum species, particularly when the pH is above 7.5. The occurrence of a sharp hydrolysis effect above this pH may be expected because of the equilibrium position of lanthanum hydroxide, indicated by21 Log [La3+] = 23-02 - 3pH When the reaction of o-p-PAN with lanthanum(II1) was investigated] only a slight colour change occurred. As the formation of the o-p-PAN chelate may be expected to occur at lower acidities than those for the a-PAN chelate, this lack of reaction may be due to prefer- ential hydrolysis of the lanthanum(II1).Unlike the absorbance curves for the manganese( 11) and zinc(I1) chelates, those shown in Fig. 5 give no direct indication of stoicheiometry. However, as the absorbance values in the presence of both excess of reagent and excess of metal ions are similar, the formation of 1 : 1 chelates can be anticipated. The neutral reagent species, HR, predominates over the pH range of chelation and so the suggested chelation reaction is KMR La3++HR + LaR2++H+ The validity of this postulate was tested by using transformation (IX) (Fig. 6).404 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, Vol. 98 Fig. 6. Comparison of experimental and theoretical absorbance curves for LaIIl(o-a-PAN) [upper graphs: (a) cxccss of metal; and (b) excess of reagent] and LaIII(o-PAY) [lower graphs: (c) excess of metal; and (d) excess of reagent]. Solid lines, experirncntal; and brokcn lines, theoretical.Concentrations : (u) CR 5.0 x M, CX 5.0 x M ; (b) CR 4.0 x 1 0 - 4 M, CM 4.0 x 10-5 M : ( c ) CR 5.0 x 10-6 M, CM 5.0 x 10-4 M ; (d) CR 5.0 x 1 0 - 4 M, CM 5.0 x 10-5 M The formation of such an LaR2+ species is in agreement with the observation of Sommer and Novotna,ll who have reported the reactions of lanthanum with PAR. These workers also found that a 1 : 1 chelate was formed and that severe hydrolysis interfered in the accurate determination of absorbances at higher pH. I l l I l l I l l 0 10 20 30 40 50 60 70 80 90 100 [HI2 x lo--* Fig. 6. Graphs of transformation (IX) in the presence of excess of lanthanum(II1): (a) o-a-PAN; and (b) o-PAP.CE 6.0 x M and CM 5-0 x MJune, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 405 Agreement between the constants listed in Table VII is not quite as good as that observed for the manganese(I1) and zinc(I1) chelates, probably because of the ease of hydrolysis of lanthanum species. In order to determine precisely the hydrolysis effect, and to verify further the proposed chelation reactions, theoretical absorbance curves were determined by using the computed complex formation constants and known values for the hydrolysis of lanthanum(II1). TABLE VII MOLAR ABSORPTIVITIES AND STABILITY CONSTANTS OF LANTHANUM(III) CHELATE Reagent Transformation EMR Transformation log K M R o-E-PAN IIIA 2.50 x 104 VIA 7.16 f 0.02 IX 2-47 x 104 XI 7-08 f 0.04 0-PAP IIIA 1.43 x 104 VIA 6-76 f 0-04 IX 1.56 x 104 XI 6-58 f 0.04 pKa, = 8.80 (0-PAP) ; 10.20 (o-a-PAN).(i) I n the pyesence of excess of Zanthanum(III)-In this instance, the absorbance is limited by the concentration of the reagent. The absorbance, A , is given by A = E M R C H R ~ ~ ~ (1 - f) - - .. .. . . (11) where cVR = molar absorptivity of MR; CHRTot = total reagent concentration; andf = fraction of reaction completed. C R - C R f=- - ‘HKTot ‘R ‘MI? .. .. .. .. . . (12) 1 - 1 + CMR/CR where CR = unreacted reagent concentration. The proposed chelation reaction is * KMR M+HR + MR+H+ .. .. . . . . (13) The conditional reaction constant, *KhR, is given by In the presence of excess of lanthanum(III), CM = CYTot, the total metal-ion concen- tration, and where the side-reaction coefficients, a1 and p, are [HR]/C, and [M]/C,, respectively.In the present study, a1 = 1 and p, if the major hydroxy-complex is La(OH),, is given by TLaI [La] + CLa(OH),I or The equilibrium constant involving hydrolysis of La3+ is given by21 La3+ + 3H,O + La(OH), + 3H+ Log [La3+] = 23.02 - 3pH Values of obtained from equation (15) were used in equation (14) to calculate Cn values. Absorbances were then obtained by using equations (1 1) and (12).406 BETTERIDGE AND JOHN: PYRIDYLAZONAPHTHOLS AND [Analyst, VOl. 98 ion concentration and are given by (ii) I n the $resence of excess of reagent-Absorbance values are now limited by the metal- .. .. .. .. . . (17) 1 f = 1 + cMR/cM *K&R = - c ~ ~ [ H l - *KMR,~~ CMCR In the presence of excess of reagent, CR = CHRTot, the total reagent concentration, and As before, a = 1 and is given by equation (15).Values of CM obtained from equation (18) were then used to calculate absorbance from equations (17) and (16). The molar absorptivity, reaction, stability and dissociation constants used for the above calculations are summarised in Table VII. Theoretical absorbance curves were drawn and compared with the experimental curves (Fig. 5). Agreement between the theoretical and experimental curves is good, which confirms that the chelation reaction suggested is correct and that the decreased absorbance at high pH is due to hydrolysis of lanthanum species. Reaction of 4-(2-pyridylazo)phenol with cobalt(II) , nickeZ(II) and copper(II)-Spot tests show that $-PAP is selective in its reactions with metal ions, and will react principally with metals of Groups VIII and IB. Cobalt(II), copper(I1) and nickel(I1) react strongly, and the chelation reactions of these ions have been investigated.Absorbance curves were obtained in the presence of both excess of reagent and excess of metal ions. Absorbance measurements were made at 548, 520 and 517 nm, these wavelengths corre- sponding to the absorbance maxima of the copper, nickel and cobalt chelates, respectively. The absorbance curves are shown in Fig. 7. These curves indicate that maximum chelation occurs at a pH greater than the pK,, value of the reagent. Also, very little chelation occurs at pH values below the pK,, value, which suggests that the neutral form of the reagent has little tendency for chelation and that chelation occurs via the anionic form.This represents an unusual type of chelation because the protons that are lost do not originate in the chelating groups themselves. Preliminary investigations of the nature of these chelates by continuous variation,l2 mole-ratio13 and s l o p e - r a t i ~ ~ ~ ~ ~ ~ methods showed that each chelate has a stoicheiometry of metal to ligand of 1 : 2. As the anionic form of the reagent is involved in chelation, then the suggested chelation reaction is M + 2R- + MR, Therefore, the application of transformations (VI) and (VII) was tried. The consistency of the calculated constants and linearity shown by the correlation coefficient in Table VIII indicated that this is the predominant reaction over the pH range 7 to 8.The decrease in absorbance observed at higher pH indicates the formation of hydroxy- chelate species : KOH MR2+OH- + MR,OH- This hydrolysis reaction is identical with that of the manganese(I1) and zinc(I1) chelates with o-p-PAN. The hydrolysis constant, KMR(OH), can therefore be calculated from identical transformations. The linearity of these transformations is shown in Fig. 8, and the hydrolysis constants were found to be 104.58 * O a o 2 and * Oeo2 for the nickel(I1) and cobalt(I1) hydroxy-species, respectively. SOLVENT-EXTRACTION STUDIES- The theory of the solvent extraction of chelates is fairly well established.24-27 If the distribution ratio, D, is defined as the concentration of the metal in the organic phase divided1.2 - 1.0 8 0-8 0.6 8 0-4 s .ff o a 0.2 6 7 8 9 1 0 6 7 8 9 1 0 1 PH Fig.7. Absorbance zlersus pH graphs for chelates of p-PAP in 5 per cent. aqueous methanol: (a) with copper(I1); (b) with nickel(I1); and ( c ) with cobalt(I1). A, excess of reagent: CM 2.5 x M, CR 1.0 X W 4 M ; B, excess of metal: Cx 1.0 x M, CR 2.5 X M ; c, excess of reagent: CM 2.5 X M, CR 1.0 x M ; D, excess of metal: CM 1.0 x M, CB 2.5 x M, CR 1.0 X lo-* M ; F, excess of metal: CM 1.0 X M ; E, excess of reagent: CM 5.0 X M, CR 5-0 X M TABLE VIII MOLAR ABSORPTIVITIES AND STABILITY CONSTANTS FOR COBALT( 11) AND NICKEL( 11) CHELATES Standard error of Condition Metal ion y = m x + c coefficient slope, e s t i m a t e rn E M R ~ x lo4 log KMR, Correlation Excess of metal ions Ni(1I) y = 0 .0 3 1 ~ + 3.820 0.994 0.00 1 1 0.088 5-23 8.95 f 0-08 W I I ) y = 0 . 1 9 8 ~ + 6-843 0.998 0.0044 0.205 2-92 7.88 f 0.07 Excess of reagent Ni(I1) y = 0.015~ $- 2.031 0-981 0.0009 0.1 46 4.92 8-69 f 0.10 Co(I1) y = 0 . 2 0 8 ~ + 3-887 0.999 0-0031 0.266 2.57 7-82 f 0.06 pKa, = 7.77.408 BETTERIDGE AND JOHN : PYRIDYLAZONAPHTHOLS AND [Analyst, VOl. 98 by the concentration of the metal in the aqueous phase, it follows, for the extraction reaction that MC+ + 2HRo + MR2(,, + 2H$ D = [MR210/([MR2]W + cM) = KDX/(l + cM/[MR21W) = KDX/(l + [Ml/p[MR21) = KDX/{l + l/(pKMR,a&D)[HRlz)) .. . . (19) In equation (19) and subsequent equations K,, and KDR are the partition coefficients of the chelate and reagent, respectively.%(D) = [R-l/c€IR = KalKa2/ { [HI2 + Kal [HI (1 + K ~ ~ v o / v w ) + Ka,Ka, 1 where CHR is the total concentration of reagent present initially in the organic phase of volume Vo and the other terms have been defined previously. It is assumed that the metal is present in the organic phase only as MR,. In the absence of side-reactions that involve the metal ion (/3 = 1) and incomplete chelate formation (D < KDx) and on the assumption that the reagent is present in the organic phase mainly as HR, this expression reduces to = K1\IR~KDXK~~[HRl~/I(~RIHla .. .. . . (20) For practical purposes, the percentage extracted, E , is often used, and is related to the distribution ratio by The pH value at E = 50 per cent. is designated as pH&, and the difference between the pH, values of two extractable chelates is a measure of their separability.If it is assumed that the partition coefficients for chelates of the same reagent are identical and there are no competing side-reactions, extraction curves (E veisus pH) for the same reagent and a series of metal ions of the same charge will have identical shapes. The relative degree of extraction at a given pH will be governed by the relative values of the formation constant ISMn,. Hence, if the values of the constants in equation (20) are known and a value of [HRIo is defined, curves of E v e w m pH can readily be calculated. PO H Fig. 8. Graphs of hydrolysis equations for cobalt(I1) and nickel(I1) chelates of p-PAP: A, cobalt(I1); and B, nickel(I1) The values of all the terms in equation (20) except KDR and KDx can be determined by the spectrophotometric methods described above.K D R must be determined independently. Therefore, in order to calculate the extraction curve, one must either assume a value of KDx or determine it independently. Two approaches seem reasonable: (i) assume that KDx = K,,, which was applied to o-PAP systems; (ii) assume that KDx is the same for similar systems, which was applied to o-or-PAN and to o-p-PAN systems, the value of lo4 being taken from previous work.14-18 The calculated curves were compared with those obtainedJune, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 409 experimentally (Fig. 9). As would be expected from the above assumptions and the additional assumption that KMRI remains constant, not all of the experimental and theoretical curves coincide. Nevertheless, they are close enough for the analyst to use the method as a guide in the choice of his experimental conditions.Fig. 9. Comparison of experimental and predicted extraction curves for the chelates of zinc and manganese: A, zinc; and B, man- ganese. Solid lines, experimental ; and broken lines, calculated. Chelates: 0, o-a-PAN; 0 , o-S-PAN; and x , o-PAP A more valuable guide is provided when there are competing reactions, /3 < 1. The simplified equation (20) must be modified by including /3 in the numerator. The competing reaction might arise from the addition of a masking agent, L, which forms a series of com- plexes with the metal ion. If the formation constants are known, it is a straightforward matter to calculate 1/p, which is given by 1 + X k , k , .. . kn[LIn. The spectrophotometric studies should reveal the presence of hydroxy-species, which can be taken into account in a similar way: n 0 l/P = 1 + KMR,(OH) [OH] + KMR,(oH), [OH], The extraction curve for the manganese(I1) - o-P-PAN system, in which the presence of hydroxy-complexes had been established spectrophotometrically, was calculated with the aid of the formation constants determined spectrophotometrically. The curve is compared with the experimentally obtained curve in Fig. 10. The agreement is satisfactory and clearly PH Fig. 10. Comparison of experimental and predicted extraction curves for Mn(o- ,%PAN),. Solid line, experimental ; and broken line, predicted410 BETTERIDGE AND JOHN: PYRIDYLAZONAPHTHOLS AND [Analyst, vOl98 demonstrates the need to select the extraction conditions with care.For more exact com- parison, the calculated and experimental pH values are given in Table IX. TABLE IX COMPARISON OF EXPERIMENTAL SOLVENT-EXTRACTION CONSTANTS WITH CONSTANTS PREDICTED BY SPECTROPHOTOMETRY Experimental - Reagent Metalion LogKf pH,,, o-a-PAN Mn(I1) 13.30 7.30 Zn(I1) 19-60 4.10 O- B-PAN Mn(I1) 16-13 7.86 Zn(I1) 22-17 4.86 O-PAP Mn(I1) 10.42 7.60 Zn(I1) 16.68 6.00 Predicted - 13.40 7-22 19.13 4.36 16.80 7-96 21.60 6.10 10.60 7-66 16.60 6-07 Log Kf PHI/, pKa, = 9.63 (o-x-PAN); 11.62 (O-p-PAN) ; 8-88 (O-PAP). KDx = lo4, assumed for each chelate. It is more usual to deduce the values of the constants in equations (19) and (20) from the experimental curve, and this procedure was followed for the manganese(I1) - o-P-PAN system so as to determine the values of KMR,(OH) and KMR,(OH),.The value of KDXKMR, was calculated from the points on the rising part of the curve of log D veysus pH and a value of 18 was calculated as a function of pH from equation (19) by using points from the descending part of the curve. fl was then examined as a function of [OH] and it was deduced that MnR2(OH),2- could not be detected, that MnR2(OH) was present and that log KMnR,(OH) had a value of 7.64. The agreement between this value and that obtained spectrophotometrically is good, and is a little surprising because of the low number of acceptable experimental points in the solven t-extracti on sys tern.CONCLUSIONS The determination of reliable stability constants is a tiresome and time-consuming procedure, which can often produce results that are of limited applicability. The most accurate results are generally agreed to be those obtained by potentiometric titrations, and these results now have the advantage of the availability of well tested computer programs28 that can be used in order to transmute the experimental points into results. However, inevitably, relatively large concentrations of reactants are used, which results in the use of non-aqueous solvents and the increased likelihood of the formation of polynuclear species. Methods based on solvent extraction and spectrophotometry have direct appeal to the analytical chemist as he may be able to set up his analytical method while measuring the stability constants.This is true of solvent extraction, although it may be difficult to obtain reliable constants without a great deal of experimentation and it is also true of the spectro- photometric method described above. The pH - absorbance curves upon which the method is based are essential in the development of the analytical procedure. These curves can also yield much information about the effect of varying the conditions, e.g., reagent concentration, pH and masking agent. In this respect, as well as in terms of accuracy, the method involving these curves is far superior to Job’s method and the related procedures that are commonly used. It has also been shown that the results can be used directly to predict solvent-extraction curves.The weakness of the method is that it depends upon extrapolation, and it requires careful experimental work and alteration of conditions so as to ensure that reliable results are obtained. We have concluded that it is just as time consuming as the other reliable methods, but we feel that because of the amount of information of direct analytical interest that is obtained, it is a very desirable method. Of the four reagents studied, o-a-PAN is clearly the most suitable as it forms highly coloured stable complexes at lower pH values than o-fl-PAN, o-PAP or @-PAP. o-fl-PAN and o-PAP form complexes at about the same pH values but the molar absorptivities of PAN complexes are greater than those of PAP complexes. Specific analytical applications will be described in subsequent papers.June, 19731 PYRIDYLAZOPHENOLS AS ANALYTICAL REAGENTS. PART I1 41 1 to one 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 28. 26. 27. 28. We are grateful to the T. and E. Williams Scholarship Fund for a maintenance grant of us (D.J.). REFERENCES Anderson, R. G., and Nickless, G., Analyst, 1967, 92, 207. Shibata, S., Analytica Chim. Acta, 1960, 23, 367. Rosotti, F. J. C., and Rosotti, H., “The Determination of Stability Constants,” McGraw-Hill, Sommer, L., Folia Fac. Sci. Natn. Univ. Purkynianae Bmo, 1964, 5, Part 1, 1. Hnilickova, M., and Sommer, L., Talanta, 1966, 13, 667. Chromy, V., and Sommer, L., Ibid., 1967, 14, 393. Sommer, L., and Hnilickova, M., Ibid., 1969, 16, 83. Sommer, L., and Ivanov, V. M., Ibid., 1967, 14, 171. Sommer, L., Ivanov, V. M., and Novotna, H., Ibid., 1967, 14, 329. Sommer, L., and Novotna, H., Ibid,, 1967, 14, 457. Vosburgh, W. C., and Cooper, G. R., J . Amer. Chem. SOL, 1941, 63, 437. Yoe, J. H., and Jones, A. L., Ind. E n g g Chem., Analyt. Edn, 1964, 16, 11. Betteridge, D., Todd, P. K., Fernando, Q., and Freiser, H., Analyt. Chem., 1963, 35, 729. Pease, €5. F., and Williams, M. B., Ibid., 1959, 31, 1044. Corsini, A., Mai-Ling Yih, I., Fernando, Q., and Freiser, H., Ibid., 1962, 34, 1090. Nakagawa, G., and Wada, H., J. Chem. SOC. Jafian, 1963, 84, 639; Chem. Abstr., 1964, 61, 1242h. Betteridge, D., Fernando, Q., and Freiser, H., Analyt. Chem., 1963, 35, 294. Anderson, R. G., and Nickless, G., Analyst, 1968, 93, 13. Pourbaix, M., “Atlas of Electrochemical Equilibria in Aqueous Solutions,” Pergamon Press, Harvey, A. E., and Manning, D. L., J . Amer. Chem. SOC., 1950, 72, 4488. Morrison, G. H., and Freiser, H. F., “Solvent Extraction in Analytical Chemistry,” John Wiley, Starf, J., “The Solvent Extraction of Metal Chelates,” Pergamon Press, Oxford, 1964. Laitinen, H. A. , “Chemical Analysis,” McGraw-Hill, New York, 1960. Marcus, Y., and Kertes, A. S., “Ion-exchange and Solvent Extraction of Metal Complexes,” Childs, C. W., Hallman, P. S., and Perrin, D. D., Talanta, 1969, 16, 1119. - Ibid., 1961, 25, 348. New York, 1961, pp. 47-51. 9 , Analytica Chim. Acta, 1967, 39, 469. -- Oxford, 1966. t , Ibid., 1952, 74, 4744. -- New York, 1962. Wiley-Interscience, New York, 1969. Received April 2nd, 1970 Accepted November 22nd, 1972